Based on promising preclinical results, a variety of TGF-beta pathway antagonists are in early phase clinical trials for the treatment of advanced cancer. However, given the complex biology of TGF-beta, the successful development of TGF-beta antagonists for cancer therapy will depend on a clear understanding of how these agents work, and the related question of how to select patients who will benefit from this type of treatment. Using a panel of 12 mouse syngeneic allograft models of metastatic breast cancer, with metastatic burden as the primary endpoint, we uncovered heterogeneous responses to TGF-beta antagonism. TGF-beta pathway blockade inhibited metastasis in some models, while having no effect on others. Importantly, in this expanded model panel, we found that TGF-beta antagonism actually stimulated metastasis in 25% of the models, suggesting that good predictive biomarkers will be crucial for safe and effective deployment of TGF-beta antagonists clinically. We have demonstrated that many of the logical or literature-suggested candidate biomarkers (eg. TGF-beta expression levels, p53 mutation status) do not reliably predict therapeutic outcome in this larger panel. Analysis of the transcriptomes of treatment-naive primary tumors allowed identification of gene signatures that were specific to the inhibitory vs the stimulatory response to TGF-beta antagonism. Applying these signatures to human breast cancer transcriptomic datasets suggested that patients with estrogen receptor-negative disease, particularly of the basal and claudin-low subtypes, might be most likely to have the desired therapeutic response to TGF-beta antagonists. Analysis of in vitro responses to TGF-beta across the mouse model panel led us to hypothesize that the undesirable stimulatory effect of TGF-beta antagonism is due to interference with inhibitory effects of TGF-beta on the cancer stem cell subpopulation. We are working to test whether this is true in vivo, using the cancer stem cell reporter developed in our companion project ZIA BC 005785, as well as other orthogonal techniques. Finally, in a new initiative to better identify and delineate the cellular targets of TGF-beta action in vivo in normal and tumor-bearing mice, we have generated a transgenic TGF-beta reporter mouse, in which activation of a novel Smad response element drives expression of a fluorescent protein. This mouse is giving important insights into patterns of TGF-beta pathway activation in the adult animal and how these are affected by pharmacologic inhibition of the pathway.